Learning with convolution and pooling operations in kernel methods

Abstract

Recent empirical work has shown that hierarchical convolutional kernels inspired by convolutional neural networks (CNNs) significantly improve the performance of kernel methods in image classification tasks. A widely accepted explanation for their success is that these architectures encode hypothesis classes that are suitable for natural images. However, understanding the precise interplay between approximation and generalization in convolutional architectures remains a challenge. In this paper, we consider the stylized setting of covariates (image pixels) uniformly distributed on the hypercube, and characterize exactly the RKHS of kernels composed of single layers of convolution, pooling, and downsampling operations. We use this characterization to compute sharp asymptotics of the generalization error for any given function in high-dimension. In particular, we quantify the gain in sample complexity brought by enforcing locality with the convolution operation and approximate translation invariance with average pooling. Notably, these results provide a precise description of how convolution and pooling operations trade off approximation with generalization power in one layer convolutional kernels.

52 pages, 6 figures

References in corpus (9)

• Stochastic Gradient Descent Optimizes Over-parameterized Deep ReLU Networks
• Enhanced Convolutional Neural Tangent Kernels
• Learning with invariances in random features and kernel models
• Kernel Alignment Risk Estimator: Risk Prediction from Training Data
• Generalization error of random features and kernel methods: hypercontractivity and kernel matrix concentration
• Locality defeats the curse of dimensionality in convolutional teacher-student scenarios
• Computational Separation Between Convolutional and Fully-Connected Networks
• Approximation and Learning with Deep Convolutional Models: a Kernel Perspective
• The Unreasonable Effectiveness of Patches in Deep Convolutional Kernels Methods